1 S/390 common I/O-Layer - command line parameters and /proc entries
2 ==================================================================
4 Command line parameters
5 -----------------------
9 Determines whether information on found devices and sensed device
10 characteristics should be shown during startup, i. e. messages of the types
11 "Detected device 4711 on subchannel 42" and "SenseID: Device 4711 reports: ...".
16 * cio_notoper_msg = yes | no
18 Determines whether messages of the type "Device 4711 became 'not operational'"
19 should be shown during startup; after startup, they will always be shown.
24 * cio_ignore = <device number> | <range of device numbers>,
25 <device number> | <range of device numbers>, ...
27 The given device numbers will be ignored by the common I/O-layer; no detection
28 and device sensing will be done on any of those devices. The subchannel to
29 which the device in question is attached will be treated as if no device was
32 An ignored device can be un-ignored later; see the "/proc entries"-section for
35 The device numbers must be given hexadecimal.
38 cio_ignore=0x23-0x42,0x4711
39 will ignore all devices with device numbers ranging from 23 to 42 and the
40 device with device number 4711, if detected.
42 By default, no devices are ignored.
45 * cio_proc_devinfo = yes | no
47 Determines whether the entries under /proc/deviceinfo/ (see below) should be
48 created. Since there are problems with systems with many devices attached, I
51 Until the problems are dealt with, default is off.
59 This entry shows information on a per-subchannel basis.
61 The data is ordered in the following way:
65 - device type/model (if applicable; if not, this is empty) and control unit
67 - whether the device is in use (i. e. a device driver has requested ownership
68 and registered an interrupt handler)
69 - path installed mask (PIM), as reflected by last store subchannel
70 - path available mask (PAM), as reflected by last store subchannel
71 - path operational mask (POM), as reflected by last store subchannel
72 - the channel path IDs (CHPIDs)
74 All fields are separated by spaces, the chpids are in blocks of four chpids.
78 Shows in subdirectories for each device some characteristics:
79 - /proc/deviceinfo/<devno>/chpids:
81 - /proc/deviceinfo/<devno>/in_use:
82 whether the device is in use
83 - /proc/deviceinfo/<devno>/sensedata:
84 the device type/model and if applicable control unit type/model of the
87 NOTE: Since the number of inodes which can be dynamically allocated by procfs
88 is limited, device entries will only be created up to a magic number of
89 devices. The kernel will utter a warning that not all entries can be
90 created. In this case, you shouldn't use "cio_proc_devinfo=yes" (see
95 Lists the ranges of device numbers which are ignored by common I/O.
97 You can un-ignore certain or all devices by piping to /proc/cio_ignore.
98 "free all" will un-ignore all ignored devices,
99 "free <devnorange>, <devnorange>, ..." will un-ignore the specified devices.
101 For example, if devices 23 to 42 and 4711 are ignored,
102 - echo free 0x30-0x32 > /proc/cio_ignore
103 will un-ignore devices 30 to 32 and will leave devices 23 to 2F, 33 to 42
105 - echo free 0x41 > /proc/cio_ignore will furthermore un-ignore device 41;
106 - echo free all > /proc/cio_ignore will un-ignore all remaining ignored
109 When a device is un-ignored, device recognition and sensing is performed and
110 the device driver will be notified if possible, so the device will become
111 available to the system.
113 You can also add ranges of devices to be ignored by piping to
114 /proc/cio_ignore; "add <devnorange>, <devnorange>, ..." will ignore the
117 Note: Already known devices cannot be ignored; this also applies to devices
118 which are gone after a machine check.
120 For example, if device abcd is already known and all other devices a000-afff
121 are not known, "echo add 0xa000-0xaccc, 0xaf00-0xafff > /proc/cio_ignore"
122 will add af00-afff to the list of ignored devices and skip a000-accc.
125 * /proc/s390dbf/cio_*/ (S/390 debug feature)
127 Some views generated by the debug feature to hold various debug outputs.
129 - /proc/s390dbf/cio_crw/sprintf
130 Messages from the processing of pending channel report words (machine check
131 handling), which will also show when CONFIG_DEBUG_CRW is defined.
133 - /proc/s390dbf/cio_msg/sprintf
134 Various debug messages from the common I/O-layer; generally, messages which
135 will also show when CONFIG_DEBUG_IO is defined.
137 - /proc/s390dbf/cio_trace/hex_ascii
138 Logs the calling of functions in the common I/O-layer and, if applicable,
139 which subchannel they were called for.
141 The level of logging can be changed to be more or less verbose by piping to
142 /proc/s390dbf/cio_*/level a number between 0 and 6; see the documentation on
143 the S/390 debug feature (Documentation/s390/s390dbf.txt) for details.
147 This entry counts how many times s390_process_IRQ has been called for each
148 CPU. This info is in /proc/interrupts on other architectures.
152 This entry will only show up if you specified CONFIG_CHSC=y during kernel
155 This entry serves a dual purpose:
157 - show which chpids are currently known to Linux and their status (online,
160 - toggling known chpids logically online/offline.
162 To toggle a known chpid logically offline, do an
163 echo off <chpid> > /proc/chpids
164 <chpid> is interpreted as hex, even if you omit the '0x'.
165 The chpid will be treated by Linux as if it were not online, which can mean
166 some devices will become unavailable.
168 You can toggle a logically offline chpid online again by
169 echo on <chpid> > /proc/chpids
170 If devices became unavailable by toggling the chpid logically offline, they
171 will become available again after you toggle the chpid online again.